Software Engineer ; AL​/ML

At my Kaarma, we’re not just leading the way in fixed ops solutions for the automotive industry—we’re redefining what’s possible for dealership service centers. Headquartered in Long Beach, California, and powered by a global team, our industry-leading SaaS platform combines communication, scheduling, and payment tools in one seamless solution that keeps dealerships and vehicle owners connected. With my Kaarma, every service interaction flows effortlessly, bringing good karma to customers and service teams.

Rooted in the principles of the Toyota Production System, we operate with precision, efficiency, and a relentless focus on continuous improvement to deliver a better experience for all. We’re looking for innovators, problem-solvers, and tech enthusiasts passionate about building solutions that people love to use. If you’re ready to make an impact in an industry ripe for change, join us at my Kaarma and help shape the future of automotive service.

• Design, develop, and test software components that enable automotive dealerships to better engage with their customers.
• Build and integrate AI/ML solutions — including developing and consuming LLM APIs, training and fine‑tuning models, and deploying them into production environments.
• Collaborate with business stakeholders and customers to translate real‑world needs into scalable technical solutions.
• Work across diverse tech stacks such as microservices, iOS/Android development, cloud infrastructure, and more.
• Communicate effectively with your team leader and product managers to set clear project goals, timelines, and deliverables.
• Drive technical excellence by speaking up, challenging tradeoffs (speed vs. reliability), and taking ownership of high‑quality outcomes.

• 1–3 years of experience developing cloud‑based software services with an understanding of scalability, performance, and reliability.
• Strong computer science fundamentals; a degree in Computer Science/Engineering or equivalent work experience is preferred (Tier 1/Tier 2 institutes like IIT, NIT, etc. are a plus, but not required).
• Skills:
  • Proficiency in at least one modern programming language (Java, Python, or C#).
  • Understanding of data structures, algorithms, and design principles.
  • Strong communication skills for working with technical and non‑technical teammates.
  • Curiosity and eagerness to learn new tools, frameworks, and architectures.
  • Operating system and networking fundamentals.
  • Data & ML Lifecycle.
• Core Concepts: Can clearly explain the purpose and importance of train/validation/test splits.
• Data Intuition: Understands “garbage in, garbage out.” Can discuss the conceptual importance of data cleaning, normalization, and augmentation.
• Evaluation: Can explain the purpose of common metrics (e.g., accuracy, precision/recall, loss). Can look at a training/validation loss curve and identify clear overfitting or under fitting.
• Core Concepts: Can intuitively explain the basic components of a neural network (neurons, layers, weights, activation functions) and their purpose.
• Training Dynamics: Understands, at a high level, how a network “learns” via loss functions, gradient descent, and back propagation.
• Practical Intuition: Can conceptually differentiate between training and inference and is familiar with core issues like overfitting and under fitting.
• Transformers.
• Motivation: Understands the motivation for the Transformer architecture (e.g., the limitations of sequential models like RNNs/CNNs for long‑range dependencies).
• Core Mechanics: Can intuitively explain the purpose of key components like self‑attention, positional encodings, and the encoder‑decoder structure.
• Attention Intuition: Grasps the basic idea of self‑attention (how a model learns which words are important to each other) and multi‑head attention (looking at a sequence in different “ways”).
• Architecture & Scaling: Understands the relationship between Transformers and LLMs and is aware of concepts like scaling laws and the pre‑training/fine‑tuning paradigm.
• Practical Pipeline: Familiar with the concepts of tokenization (how text is processed) and prompt‑based interaction.
• Application Concepts: Aware…